The present invention relates to a thin film transistor (TFT) having optical-electrical conversion function or an image sensor, and producing method of such a transistor. In particular, the present invention relates to such a transistor, or an image sensor, which is excellent in productivity or producing yield rate.
Conventionally, a photo-diode, and/or a photo-conductor type cell has been used as an optical-electrical converter or an image sensor in a facsimile device and/or a copy machine et al. In those devices, amorphous silicon, single crystal silicon, C.sub.d S, C.sub.d S.sub.e, C.sub.d S-S.sub.e have been used.
FIG. 7 shows a prior solid state image sensor. FIG. 7A is a circuit diagram of a sensor module, and FIG. 7B is an enlarged cross section of the portion enclosed by dotted line in FIG. 7A.
In FIG. 7A, the numeral 91 is a thin film transistor (TFT), 92 is a photo-diode, and 93 is a shift register, which is coupled with the photo-diode for operating the same. The symbols V.sub.DD, V.sub.SS, CL, V.sub.BB, GND et al show terminals for input, output, and/or power supply.
As shown in FIG. 7B, the thin film transistor 91 and the photo-diode 92 are produced on a single substrate 94 made of quartz. The transistor 91 comprises a poly-silicon layer 95, a gate insulation film 96, and a gate electrode 97, and the photo-diode 92 comprises an amorphous silicon layer 100, an amorphous silicon layer 101 doped with boron, and ITO (indium tin oxide) layer 102. In FIG. 7B, 98 is an oxide layer, 99 is an electrode, and 103 is a passivation layer.
A prior image sensor has, separately, an image sensor portion made of a photo-diode 92, and a drive circuit portion made of a thin film transistor 91, and those portions are produced separately.
As for an image sensor, not only an photo-diode of FIG. 7, but also a photo-transistor using amorphous silicon, and/or a multi-chip type image sensor using single crystal silicon is possible.
However, a prior solid state image sensor shown in FIG. 7 has the disadvantage that an image sensor portion and a circuit portion for operating said image sensor portion are produced separately. Therefore, the producing steps must include the steps for producing an image sensor portion, the steps for producing the related circuit portion, and the steps for coupling those two portions. Therefore, the producing steps are complicated, and producing yield rate decreases, and the producing cost increases.
In addition, as a photo-diode for an image sensor has no function for amplification, a separate amplifier circuit must be attached.
Another prior image sensor, for instance, a photo-transistor using amorphous silicon has the disadvantage that optical response is slow, and it is not used for high speed facsimile system. It has further the problem for stability of layers, and operational reliability.
Another prior image sensor, like, multi-chip type image sensor using crystal silicon has the disadvantage that characteristics of each chip are not uniform, and therefore, an image sensor for intermediate darkness is difficult. In addition, the coupling between chips is difficult, and therefore, it is impossible to provide excellent picture. Further, it must have an selfoc lens when it is used as a contact type image sensor.
Further, in practical application of an image sensor for a facsimile which has a plurality of image sensors, the electrical characteristics of each image sensor are not uniform due to ununiformness of mobility and/or threshold voltage of each image sensor, and therefore, output potential of each image sensor relating to each dot of facsimile signal at dark condition is not uniform.
If dark output is not uniform, the measure of illumination is difficult, since dark output is reference output. Therefore, conventionally, dark output for each pixel is stored in a ROM (read only memory), and a bright output of an image sensor is compared with a dark output which is stored in the ROM, in order to determine whether it is dark or bright.
Therefore, conventionally, dark output of each pixel of a series of image sensors is measured and the measured value is stored in the ROM. Therefore, an image sensor for a facsimile system is very expensive.